IK-SVD: Dictionary Learning for Spatial Big Data via Incremental Atom Update

Big Data, a large and complex collection of datasets characterized by four V's (volume, variety, veracity, and velocity), is difficult to deal with using traditional data processing algorithms and models. A proposed dictionary learning algorithm, which extends the classical method that uses the K-means and Singular Value Decomposition (K-SVD) algorithm by incrementally, updating atoms, will ably represent the spatiotemporal remote sensing of Big Data and do so both efficiently and sparsely.